Semantics Turkish Non-Past

SEMANTICS of the TURKISH NON-PAST

A Senior Essay by Ryan Caro

Advisor: Ashwini Deo Department of Linguistics, Yale University

Spring 2012

“Nereye gidece ğini bilmiyorsan, herhangi bir yol seni oraya götürür.”

Caro Semantics of the Turkish Non-Past

Table of Contents

Abstract 2 Acknowledgements 3 1. Introduction 4 2. Previous models 4 2.1 Haspelmath 2.2 Yava ş 3. Uses of the Progressive 6 3.1 Progressive 3.2 Habitual 3.3 Bouletic future 3.4 Stative verbs 3.5 Changes in the Progressive since Yava ş 4. Uses of the Aorist 14 4.1 Characteristic/Generic 4.2 Inertial Future 4.3 Conditional 4.4 Historical Present 4.5 Precative 4.6 Performative 5. Uses of the Future 23 5.1 Future Time Reference 5.2 Performative 6. Formalizing the Distinction between the Aorist and Progressive 25 7. Formalizing the Performative 27 7.1 The Absence of a Progressive Performative 7.2 The Performative Future 7.3 The Performative Aorist 8. Conclusion 29 References 31

Caro Semantics of the Turkish Non-Past

Abstract

In Turkish, three morphemes expressing tense/aspect cover the semantic categories of the present and future: the Progressive, Aorist, and Future. Though both Haspelmath (2003) and Yavas (1982) make claims about the division of labor between these three grammatical morphemes, their precise nature is not properly explored due to a paucity of data. In this essay, I review these analyses and then, in detail, recount the specific semantic fields which are correlated with each morpheme. In particular, I focus on all three grams with Future Time Reference; I also look at the performative uses of the Aorist and Future, and establish why the Progressive cannot be used performatively. Finally, I use this data, along with the history of each morpheme, to posit formal representations for them.

Caro Semantics of the Turkish Non-Past

Acknowledgements

I owe deepest gratitude to my advisor, Ashwini Deo of the Department of Linguistics at Yale. In addition to bringing me through the process of writing this essay, she has provided invaluable advice to guide me through the end of my career as a student. I owe her a debt beyond these words. I am also grateful to Raffaella Zanuttini and Jelena Kripokavic of Yale’s Linguistics Department, in addition to the other linguistics majors, for their guidance and patience through the many versions of this essay. I would like to acknowledge my main informant, Sezgin Çetinkaya, for improving my understanding of the Turkish language and allowing me to subject him to endless questions about verb forms. Te şekkür ediyorum! Additional thanks is owed to J. Errington of the Anthropology Department at Yale and Gregory Guy of the Linguistics Department of New York University for their guidance in my first planned senior essay on the history of the return of the second-person singular pronoun tu to the Portuguese of Rio de Janeiro. Though that project never came to fruition, I am thankful for their assistance. This project and, indeed, my entire career at Yale would not have been possible without the love and support of my parents, relatives, friends, and colleagues, both at Yale and elsewhere. I would especially like to thank Dean Camille Lizarribar of Ezra Stiles College for her continued encouragement and support through what has been, by all accounts, a difficult year. Finally, I would like to dedicate this essay, however unworthy, to my grandmother, Marie Caro, who passed away in April 2012. Though she did not live to see me graduate, this essay represents the culmination of the accomplishments that made her so proud.

This essay is, in part, an expansion of a term paper written for Linguistics 265: Semantic Change with Professor Deo in the fall of 2011.

Caro Semantics of the Turkish Non-Past

1. Introduction

The Progressive, Aorist, and Future in Turkish are all ‘grammatical morphemes’ (or ‘grams’) in the sense laid out by Bybee and Dahl (1989): language-specific categories with a certain form which encompasses a number of specific meanings. When used alone, these three morphemes cover all categories of tense and aspect which do not relate to the past. This essay deals exclusively with the Progressive, Aorist, and Future used alone, rather than in combination with the two past morphemes in Turkish, the Past –DI and Reported Past -mI ş. In this essay, I first present previous accounts of the semantic distinctions between these three morphemes. I go on to present data for different semantic categories, demonstrating which of the grams can be associated with which category. Finally, I analyze the ways in which these categories are connected to create a semantic map. A note about style: In order to maintain a contrast, I have chosen to use initial majuscules for names of language-specific grams (e.g. Aorist, a system first proposed by Comrie 1976), while reserving SMALL CAPS for semantic categories. In the cases where I discuss cross- linguistic “gram-types” like the canonical “Present” I use quotation marks. In addition, I follow standard conventions in Turkish linguistics by using capital letters to represent underspecified vowels which undergo vowel harmony. A represents a low, unrounded vowel, unspecified for backness, which can be realized as a or e. I represents a high vowel unspecified for backness and rounding, which can be realized as ı [ɯ], i, u, or ü [y]. For simplicity’s sake, I have chosen to maintain Turkish orthography rather than use the International Phonetic Alphabet. The following letters differ substantially from their IPA symbols: c [d ʒ], ç [t ʃ], ğ [:] (lengthens preceding vowel), ı [ɯ], j [ʒ], ö [ø], ş [ʃ], ü [y], y [j].

2. Previous models

2.1 Haspelmath

Haspelmath (2003) discusses the benefits of semantic maps, his solution to the problem of multifunctionality in grammar. As demonstrated in (1), one grammatical morpheme can encompass several meanings:

(1) English Past Tense (Haspelmath 2003) a. Goethe wrote a poem every day. (PAST HABITUAL )

Caro Semantics of the Turkish Non-Past

b. Goethe wrote Faust in 1808. (PAST PERFECTIVE ) c. If she wrote to me tomorrow, I would reply in time. (HYPOTHETICAL )

Semantic maps provide a solution to the issue of determining what, precisely, is the meaning of a particular gram by laying out the cross-linguistic categories with which language-specific grams are associated. Each gram can then be laid out over these semantic areas, which should form a contiguous area on the map, demonstrating the extent of its meaning. Thus semantic maps can delineate the precise meaning of a gram without resorting to a vague generalization or rampant polysemy. Haspelmath’s semantic maps also have historical consequences; he purports that a gram changes diachronically to encompass a new meaning only if that new meaning is already connected to one of its previous meanings. Though this process still leaves every gram with a contiguous area on the semantic map, the development of a new gram with only one, central meaning might leave the semantic area of an older gram with a hole in the middle, creating, in Haspelmath’s words, a “doughnut.” Haspelmath alleges that this process took place in Turkish, with the Progressive and Aorist. Historically, the Aorist was a “Present” which could express events occurring at speech time, and has retained such uses in closely related languages like Azeri (also known as Azerbaijani, Johanson 1989). However, present-day Turkish has a Progressive which accommodates these events, while the Aorist is disallowed for them. Still, the Aorist is used for other categories of events. Here, then, is Haspelmath’s “doughnut”: while still encompassing the

HABITUAL and FUTURE , it can no longer be used to express PROGRESSIVE meanings, the bridge between HABITUAL and FUTURE on his semantic map (Figure 1).

Figure 1 . Haspelmath’s proposed semantic map of the Turkish Aorist

In this paper, however, I argue that not only is the Aorist unacceptable for many of the situations for which Haspelmath (2003) claims it is used, but also that the situations for which it is acceptable form a contiguous semantic area which represent those semantic categories which make statements about the properties of the world.

Caro Semantics of the Turkish Non-Past

2.2 Yava ş

Haspelmath (2003) claims that the Aorist in Turkish covers the categories of HABITUAL and FUTURE , basing his information on Yava ş (1982). Yava ş, however, emphatically denies that these are the uses of the Aorist. Rather, he argues, the Aorist is compatible with only characteristics of an individual, while temporally-bounded habits are left to the Progressive. In addition, he argues that the Aorist is only compatible with Future Time Reference when the speaker is relatively uncertain about the event, not with all instances of FUTURE . Yava ş characterizes the Aorist as describing “what is typical, normal, or inherent to an entity or to a situation.” This comes closer to an accurate description of the Aorist, but his article gives only a vague formalization of the precise nature of the gram. In the following sections, then, I present and analyze data which gives a more comprehensive description of the semantics of the Aorist, in addition to the Future and, first, the Progressive.

3. Uses of the Progressive

The Progressive in Turkish historically derives from its original, and still central, meaning of PROGRESSIVE , the semantic area that deals with events in progress at reference time. Its form is the suffix –Iyor , which, because it is derived from the old verb yorı- ‘walk, go’ (cf. modern Turkish yürü- ‘walk’), does not undergo the expected vowel harmony in the second vowel. From this core, the Progressive has progressed to encompass many other semantic areas, to the point that Dahl (1985) analyzes the Progressive as an “Imperfective.”

3.1 Progressive

The Turkish Progressive is used to describe events in progress at reference time. When not combined with any past tense morphemes, reference and speech time are coterminous, so events in progress at reference time are also events in progress at the time of speech, as in (2-4).

(2) Çocuk-lar bahçe-de oyn-uyor-lar Child-PL garden-LOC play-PROG -PL ‘The children are playing in the yard.’

Caro Semantics of the Turkish Non-Past

(3) Sahil Güvenlik: Dizi-nin Yıldız-lar-ı şimdi ne-ler yap-ıyor 1 Coast Guard Soap.opera-GEN star-PL -POSS now what-PL do-PROG ‘Baywatch: What are the Soap Opera Stars doing now?’

(4) Yol-un son-u gör-ün-üyor 2 Road-GEN end-POSS see-REFL -PROG ‘The end of the road is in sight.’

These are all prototypical cases of PROGRESSIVE event descriptions, in which the time of the event contains the speech and reference times.

3.2 Habitual

There is no doubt that the Progressive has taken over the meaning of PROGRESSIVE from the Aorist in the seventeenth and eighteenth centuries (Johanson 2009). However, a closer look at the article Haspelmath cites shows that the Aorist no longer conveys HABITUAL meaning, nor

many instances of FUTURE . In fact, Yava ş (1982) claims that the semantic category encoded by the Aorist is precisely not the HABITUAL , demonstrated in (5), which allows the Progressive but not the Aorist.

(5) Bu sene bayan-lar daha büyük şapka giy-iyor-lar / *giy-er-ler (Yava ş 1982) This year lady-PL more big hat wear-PROG -PL / *wear-AOR -PL ‘Women are wearing larger hats this year’

Indeed, any habit, so long as it is temporally bounded, must take the Progressive, while the Aorist is disallowed.

(6) Bebek bu gün-ler-de çok öksür-üyor / *öksür-ür (Yava ş 1982) Baby these day-PL -LOC a.lot caugh-PROG / *cough-AOR ‘The baby is coughing a lot these days.’

(7) Türk sinema-sı son yıl-lar-da güzel film-ler yap-ıyor mu siz-ce? 3 Turkish cinema-POSS last year-PL -LOC nice film-PL make-PROG -Q you-language ‘In your opinion, has the Turkish film industry been making good films in recent years?’

1 The title of an article: http://starlounge.tr.msn.com/gallery.aspx?cp-documentid=160669621 2 The title of a popular song by İbrahim Tatlıses: http://www.youtube.com/watch?v=Mv25BB5rINs 3 A questions posted on Formspring: http://www.formspring.me/r/turk-sinemas-son-y-llarda-guzel-filmler-yap-yor- mu-sizce/230725965364337736

Caro Semantics of the Turkish Non-Past

(8) Bu hafta balık y-iyor-um 4 This week fish eat-PROG -1SG ‘This week I am eating fish.’

(9) Gerçek ad-ı LuAyne Barber, sahne ad-ı Lulu ol-an kadın 2006-dan real name-POSS LuAyne Barber stage name-POSS Lulu be-SBJP woman 2006-ABL beri dans ed-iyor 5 6 since dance-PROG ‘The woman, whose real name is LuAyne Barber and stage name is Lulu, has been dancing since 2006’

These examples all share important characteristics which distinguish them from being

instances of the Progressive used in its simple PROGRESSIVE meaning. First, all are iterative events; the baby in (6) has certainly not been continuously coughing for days on end, nor has the woman writing sentence (8) been continuously shoveling fish into her mouth. Example (5) could also be uttered at night, when all women have retired to bed and taken off their hats. In these cases the event may not be in progress at speech time, meaning that they are not simple cases of the PROGRESSIVE . In addition, all are specifically temporally bounded with an adverbial expressing time: ‘this year,’ ‘these days,’ etc. Though the time period can be extensive (years, or even longer) this delimiting of the time period in which the habitual action takes place is important. Even when the temporal period is not explicitly stated, it can be inferred from the context, as in (10), which is similar to (6):

(10) Yeni do ğ-mu ş bebe ğ-im çok ağl-ıyor, ne yap-malı-yım? 7 New be.born-PERF baby-1SG a.lot cry-PROG what do-NEC -1SG My newborn baby has been crying a lot; what should I do?

(10), a question sent to a video blogger, demonstrates the HABITUAL meaning, despite the lack of an explicitly define time period.

3.3 Bouletic future

Haspelmath does not take into account the entirety of Yava ş’s discussion of the FUTURE when claiming that the Aorist in Turkish still covers that semantic category. In fact, Yava ş

4 From a diet forum on eating meat: http://www.kadinlarkulubu.com/archive/t-316905-p-6.html 5 The caption of a photograph from Milliyet, a popular newspaper: http://www.milliyet.com.tr/fotogaleri/44939- yasam-hobisi-direk-dansi/3 6 SBJP for “subject particle” is another gloss borrowed from Kornfilt (1997) 7 A question asked to a parenting expert: http://www.uzmantv.com/yeni-dogmus-bebegim-cok-agliyor-ne- yapmaliyim

Caro Semantics of the Turkish Non-Past

argues that the Progressive can be used for an entire category of statements with Future Time Reference, namely those with a high level of certainty for the speaker, as in (11):

(11) Karar ver-di-k, sene-ye Amazon-a gid-iyor-uz (Yava ş 1982) Decision give-PST -1PL year-DAT Amazon-DAT go-PROG -1PL ‘We’ve decided; next year we are going to the Amazon.’

There are, however, statements with Future Time Reference which are entirely certain but which cannot use the Progressive; rather, as in (12), they are allowed only with the Future (12b) and the Aorist (12c).

(12) a. * Yarın güne ş saat 5:13-te do ğ-uyor 8 Tomorrow sun hour 5:13-LOC be.born-PROG ‘The sun is rising at 5:13 tomorrow.’

b. Yarın güne ş saat 5:13-te do ğ-acak Tomorrow sun hour 5:13-LOC be.born-FUT ‘The sun will rise at 5:13 tomorrow.’

c. Yarın güne ş saat 5:13-te do ğ-ar Tomorrow sun hour 5:13-LOC be.born-AOR ‘The sun rises at 5:13 tomorrow.’ 9

Yava ş’s analysis, though it contradicts Haspelmath’s, does not account for this data. This presents a problem: why is something as predictable as the time of sunrise acceptable only with the Future and Aorist, the grams which he associates with lower degrees of certainty, while the Progressive cannot be used? An explanation lies in Copley’s (2009) notion of a bouletic director: the Progressive in both English and, as shown here, Turkish, is acceptable only if there is a director, which Copley defines as an individual or group of individuals who direct an event and are committed to that event’s occurrence. This contrasts with inertial orderings, which lack a bouletic director and cannot, in Turkish, be expressed with the Progressive. Copley makes the distinction clear in (13), an example from English:

8 A Google search for “yarın güne ş * do ğuyor” returns a single result from Twitter, while the same sentence with the Future (as in 10b) returns over 92,000 hits, whereas the Aorist (as in 10c) returns 82,000. 9 The fact that the English Simple Present is acceptable here, while the English Progressive is not, despite both being futurates, is examined more closely in Copley (2009), but not immediately relevant to the case of Turkish.

Caro Semantics of the Turkish Non-Past

(13) I think that I will go to Harvard Square tomorrow… (Copley 2009) a. …I've been meaning to get some shopping done. (Bouletic) b. …that's just the kind of thing I might do. (Inertial)

In the bouletic interpretation (13a), the speaker will go to Harvard square because she is the director; she is committed to the event of her movement to Harvard Square occurring. (13b), on the other hand, represents an inertial ordering because the event is not predicated on the desires of the speaker (or another director) but rather on the properties of the world. In all inertial worlds (ones in which the world proceeds in accordance with its properties), therefore, the speaker will go to Harvard Square the following day. Further examples make it clear that only a bouletic reading is possible with the Progressive:

(14) Gelecek sene-ye tur-a çık-ıyor-uz 10 Future year-DAT tour-DAT go.out-PROG -1PL ‘Next year we are going on tour’

(15) Hafta-ya spor-a ba şl-ıyor-um 11 Week-DAT sport-DAT begin-PROG -1SG ‘Next week I am starting to exercise.’

All events with bouletic directors can use the Progressive, but any event without a bouletic director (such as a sunrise) is not acceptable in the Progressive, no matter how certain it is.

3.4 Stative verbs

Stative verbs are those which express states, which, unlike non-stative events (the category which covers accomplishments, activities, and achievements) “are not processes going on in time.” In English, stative verbs are felicitous with the Simple Present, as in (16a, 17a, 18a), but cannot generally be expressed with the Progressive be … –ing , as in (16b, 17b, 18b):

(16) a. This book has three sections.

b. *This book is having three sections.

(17) a. I love Paris in the springtime.

10 Quoted from an article about Rihanna returning to the stage: http://www.acunn.com/haber/rihanna-sahnelere-geri- donuyor/7151 11 From a dieting forum: http://www.idealdiyet.com/emziren-anne-diyeti-ile-zayiflama

Caro Semantics of the Turkish Non-Past

b. *I am loving Paris in the springtime.

(18) a. Seo-yun knows the answer.

b. *Seo-yun is knowing the answer.

In Turkish, however, stative verbs prefer the Progressive to the Aorist.

(19) O-nu tan-ıyor-um She-ACC know-PROG -1SG . ‘I know her.’

(20) O-ndan ho şlan-ıyor-sun He-ABL like-PROG -2SG ‘You like him.’

(21) Utan-ıyor Feel.embarrassed-PROG ‘S/he feels embarrassed.’

(22) Tarkan-ın yeni şarkı-sı-nın güzel ol-du ğ-u-nu dü şün-üyor-um 12 Tarkan-GEN new song-POSS -GEN nice be-FNOMNFUT -3SG -ACC think-PROG -1SG ‘I think that Tarkan’s new song is nice.’

Analogous sentences with the Aorist all have event readings compatible with other semantic areas of the Aorist, such as the CONDITIONAL (23, discussed in 4.3, below):

(23) Şimdi Fatih kalk-sa mezar-ı-ndan, ne ben on-u tanı-r-ım ne Now Conqueror rise-COND grave-POSS -ABL neither I he-ACC know-AOR -1SG nor o ben-i tanı-r13 he I-ACC know-AOR ‘If the Conqueror were to rise from his grave, I would not recognize him, nor would he recognize me.’

Here the verb tanı- ‘to know’ is interpreted as the event verb ‘to recognize’ when used with the Aorist, because it can only be stative with the Progressive. According to Vendler (1957), stative verbs cannot take progressive aspect, but they are expressed almost exclusively by the Progressive in Turkish. The Progressive, therefore, must have developed into a gram much like

the English Present, which expresses HABITUAL , BOULETIC FUTURE , and STATIVE PRESENT , but

12 Here I borrow Kornfilt’s (1997) gloss of non-future factive nominal, one of the most common ways of creating an embedded clause in Turkish. 13 From the lyrics to a nationalist Turkish anthem: http://www.youtube.com/watch?v=dpcvRZOf5BU

Caro Semantics of the Turkish Non-Past

which, unlike the English Present, has maintained its use to mark PROGRESSIVE . This is not

surprising given Haspelmath’s (2003) close association of HABITUAL -PROGRESSIVE -FUTURE on his semantic map, as well as Bybee et al.’s (1994) observation that grams’ meanings are a product of their historical development. Taken together, this grouping of semantic categories implies that the Turkish Progressive is best analyzed as an “Imperfective,” a conclusion which Dahl (1985) also draws.

3.5 Changes in the Progressive since Yava ş

The following section deals with the semantic categories covered by the Aorist, but there is one final phenomenon of note in discussing the Progressive. Perhaps unsurprisingly, given its history of coopting semantic categories previously covered by the Aorist, the Progressive seems to be encroaching on the CHARACTERISTIC /GENERIC , discussed in 0.0, below. Yava ş had reported a sharp difference between the two, with the Progressive required for the temporally bounded

HABITUAL , while the Aorist was used for the unbounded CHARACTERISTIC , a distinction demonstrated in (24):

(Yava ş 1982) (24) Baba-m erken yat-ar ama son iki sene-dir geç yat-ıyor Father-1SG early lie.down-AOR but last two year-for late lie.down-PROG ‘My father goes to bed early, but these last two years he has been going to bed late’

In the intervening three decades between his paper and this one, however, the requirement that an event be temporally bounded for it to be expressed with the Progressive seems to have been lost.

(25) a. Baba-m erken yat-ıyor Father-1SG early lie.down-PROG ‘My father goes to bed early.’

b. Baba-m hep / her zaman erken yat-ıyor Father-1SG always / every time early lie.down-PROG ‘My father always goes to bed early.’

c. Baba-m her gün erken yat-ıyor Father-1SG every day early lie.down-PROG ‘My father goes to bed early every day.’

Caro Semantics of the Turkish Non-Past

These statements, then, are available in both the Progressive and Aorist. When the Aorist is used they are interpreted as characteristics of an individual, as discussed in 4.1 below, whereas when the Progressive is used they are more like temporally unbounded HABITUALS . Sentence (24), however, cannot be rewritten in the Progressive, as in (26):

(26) *Baba-m erken yat-ıyor ama son iki sene-dir geç yat-ıyor Father-1SG early lie.down-PROG but last two year-for late lie.down-PROG ‘My father goes to bed early, but these last two years he has been going to bed late’

This sentence seems contradictory because both statements use the Progressive, suggesting that both actions are in progress at the time of speech; it is clear, however, that if the father has been going to bed late it is not possible for him to also be going to bed early. This example, therefore, usefully shows the continued distinction between the Progressive and the Aorist in the contrast between the acceptable and unacceptable ((24) and (26), respectively) sentences above: in order for the Progressive to be used, the action must be ongoing at speech time, not merely characteristic of that individual. (27) provides another examples, in which (27a) and (27b) demonstrate that both the Progressive and Aorist can be used for an unbounded

HABITUAL and CHARACTERISTIC , respectively, but the Progressive cannot be used for a

CHARACTERISTIC , as in (27c).

(27) a. Kasab-ım iyi et sat-ıyor butcher-1SG good meat sell-PROG ‘My butcher is selling good meat’

b. Kasab-ım iyi et sat-ar butcher-1SG good meat sell-AOR ‘My butcher sells good meat’

c. *Kasab-ım normal-de iyi et sat-ıyor ama bu gün-ler-de kötü gel-iyor butcher-1SG normal-LOC good meat sell-AOR but this day-PL -LOC bad come-PROG Intended reading: ‘My butcher sells good meat but it has been bad recently’

To encompass this change over the past three decades, the only change to our understanding of the HABITUAL that need be made is to point out that it is applicable to any action which occurs iteratively over a time period which includes speech time, whether temporally bound or not. This means that there are a multitude of situations which are both

Caro Semantics of the Turkish Non-Past

HABITUAL and CHARACTERISTIC /GENERIC (i.e. any situation in which an individual has been acting out a characteristic at speech time), but this analysis is advantageous in that the understanding of the CHARACTERISTIC /GENERIC in Section 4.1 need not be changed at all.

4. Uses of the Aorist

The term aorist , derived from Greek aoristos ‘indefinite’ generally refers, in the Indo- European tradition, to perfective markers which were deemed the simplest form of a verb (Lewis 2000). In Turkish, however, the name “Aorist” has been applied to an affix which, at least when used without any other tense, aspect, or modality markers, refers exclusively to non-past situations. Morphologically, the Turkish Aorist has three allomorphs: -r after vowel-final roots, - Ar after most monosyllabic roots, and –Ir after polysyllabic roots and about a dozen monosyllabic roots, most of which end in r or l.

Historically, the Aorist was a present tense which subsumed the categories of HABITUAL ,

PROGRESSIVE , and BOULETIC FUTURE , which have been taken over by the Progressive. The semantic areas with which it can still be used are outlined in this section.

4.1 Characteristic/Generic

Section 3.2, above, demonstrated that the Progressive covers the HABITUAL , while 3.5

shows that the HABITUAL can be unbounded temporally. Section 3.2 stated that this contrasted

with the CHARACTERISTIC /GENERIC , a semantic area with which this section deals. As Yava ş demonstrated with example (24) (repeated here as (28)), there is a sharp distinction between iterative actions which can be described using the Progressive and those that can be described using the Aorist. The Aorist ones, which I term CHARACTERISTIC , are different

from HABITUAL because CHARACTERISTIC marks the properties of an individual, while HABITUAL expresses the feature of a specific period of time.

(Yava ş 1982) (28) Baba-m erken yat-ar ama son iki sene-dir geç yat-ıyor Father-1SG early lie.down-AOR but last two year-for late lie.down-PROG ‘My father goes to bed early, but these last two years he has been going to bed late’

Thus, in (28), the Aorist has a CHARACTERISTIC reading, that the father is the type of

person who goes to bed early, while the Progressive has a HABITUAL reading, that the period of

Caro Semantics of the Turkish Non-Past

these past two years is marked by the fact that the father has been going to bed late. The statement which uses the Aorist must be expressing CHARACTERISTIC precisely because the action described (going to bed early) does not hold at the time of speech. In other words,

because the Progressive statement is expressing HABITUAL , the Aorist statement cannot do so because if it did the two habits would be in conflict with one another. Yava ş points out that the distinction is still clear for utterances which do not involve a time expression, as in (29a, b).

(29) a. Benim kasab-ım iyi et sat-ar (Yava ş 1982) My butcher-1SG good meat sell-AOR ‘My butcher sells good meat.’

b. Benim kasab-ım iyi et sat-iyor My butcher-1SG good meat sell-PROG ‘My butcher is selling good meat.’

(30) Ben yap-ar-ım da Tayfur yap-maz 14 I do-AOR -1SG also Tayfur do-NEG .AOR ‘I would do it, but Tayfur would not’

As in (28), the sentence with the Aorist (29a) can hold true even at a time when the butcher is selling meat of particularly poor quality, whereas (29b), which uses the Progressive and expresses HABITUAL , cannot. Example (30) is an interesting case: this is a quote from Süleyman Seba, honorary president of Be şikta ş Gymnastics Club, whose soccer team is one of Turkey’s most popular. Seba’s nephew, Tayfur Havutçu, who is the team’s technical director, was accused of being involved in a game rigging scandal. With this utterance Seba is registering his disbelief to the press; despite the fact that no one has accused him of being involved, he can use the Aorist to say, essentially “It is more characteristic of me [to rig a game] than it is of Tayfur.” As in (29a), where selling good meat is a characteristic of the butcher in question whether or not he is selling good meat at the time of speech, so in (30) can Seba use the Aorist to make a claim about what is characteristic of him and his nephew without regard to what has actually taken place.

Yava ş ties this CHARACTERISTIC meaning to a more general use of the Aorist, which is its use as “timeless tense” (Menges 1968) which expresses facts about the world, as in (32, 33).

14 A quote from CNN Turkey: http://www.cnnturk.com/2011/spor/futbol/07/15/seba.ben.yaparim.da.tayfur.yapmaz /623064.0/index.html

Caro Semantics of the Turkish Non-Past

(32) İki kere iki dört ed-er (Yava ş 1982) Two times two four make-AOR ‘Two times two make four’

(33) Dünya güne ş-in etraf-ı-nda dön-er Earth sun-GEN around-POSS -LOC rotate-AOR ‘The earth revolves around the sun’

Examining these two meaning together, the CHARACTERISTIC and Menges’s “timeless tense” use, which I will call GENERIC , Yava ş argues that the Aorist “characterizes what is typical, normal, or even inherent to a situation.” This analysis effectively explains the connection between the two groups: CHARACTERISTIC , such as (29a), and GENERIC , which includes statements like (32, 33), as well as those in (34, 35), which have a generic singular.

(34) Ku ş uç-ar Bird fly-AOR ‘Birds fly’

(35) Ördek vakvakla-r Duck quack-AOR ‘Ducks quack.’

4.2 Inertial Future

Section 3.3 demonstrated that the Progressive with Future Time Reference is only felicitous with bouletic orderings, rather than inertial ones. Inertial orderings, however, are possible with the Aorist; indeed, the Aorist can reference future time only in inertial cases. The distinction is shown in (36a, b):

(36) a. Karar ver-di-k, sene-ye Amazon-a gid-iyor-uz (Yava ş 1982) Decision give-PST -1PL year-DAT Amazon-DAT go-PROG -1PL ‘We’ve decided, next year we’ll go to the Amazon.’ (Bouletic, Progressive)

b. Amazon çok enteresan ol-malı; henüz gör-me-di-k, herhalde sene-ye gid-er-iz Amazon very interesting be-NEC yet see-NEG -PST -1PL probably year-DAT go-AOR -1PL ‘The Amazon must be very interesting; we haven’t seen it yet; we’ll probably go next year.’ (Inertial, Aorist)

(36a) uses the Progressive not because of its certainty, as argued by Yava ş, but because the deciders are the bouletic directors of the event; they have committed themselves to a trip to the Amazon. In (36b), however, there has been no such commitment; rather, the speaker is

Caro Semantics of the Turkish Non-Past

merely stating that given the properties of the world (that the Amazon is interesting, and that they are the type of people that see new, interesting things), it is expected that she and others will visit the Amazon, an inertial conjecture. Further examples demonstrate that the Aorist is compatible only in cases without a bouletic director (or, at least, in which the speaker does not believe there to be a director), in which the future action depends on characteristics already present in the world:

(37) Avrupa ekonomi-si gelecek sene düzel-me-ye ba şla-r15 Europe economy-POSS future year recover-FNOM -DAT begin-AOR ‘The European economy will begin to recover next year.’

(38) Belki bu yaz um-du ğ-umuz-dan iyi geç-er 16 Maybe this summer hope-FNOMNFUT -1PL -ABL good pass-AOR ‘Maybe this summer will go better than we hope.’

An additional proof of the lack of a director in Aorist utterances is the presence of the phrase “ İnşallah” in sentences with the Aorist with Future Time Reference. “ İnşallah” comes in ʃɑʔ ɑl: ɑ:] and literally means ‘If God wills,’ though it is used by] إن ء ﷲ from the Arabic Turks both secular and religious with a more diluted meaning closer to the (admittedly much derided) American ‘hopefully.’ “ İnşallah” implies lack of human control over the situation, a scenario very closely linked to lack of bouletic direction. Indeed, a search for “ İnşallah” with future time adverbials reveals many examples of the Aorist with inertial ordering, but none of the Progressive:

(39) İnşallah gelecek-te daha iyi ol-ur 17 İnşallah future-LOC more good be-AOR ‘God willing, it will be better in the future.’

(40) İnşallah sene-ye yaz-ar-ım 18 İnşallah year-DAT write-AOR -1SG ‘God willing, I will write it next year.’

15 The title of an article from the economics section of Zaman, a popular newspaper: http://www.zaman.com.tr/ haber.do?haberno=735246 16 From an editorial on a regional news site: http://www.iskenderunhaber.com/2011/06/uzatin-elinizi/# 17 A comment from a Youtube video showing a toy plane crashing during an unsuccessful take-off: http://www.youtube.com/all_comments?v=SIXEEpc7OHw 18 A facebook group: https://www.facebook.com/pages/in%C5%9Fallah-seneye- yazar%C4%B1m/134058333365913

Caro Semantics of the Turkish Non-Past

(41) İnşallah daha sonra iş-im-e yara-yan bir-şey-ler bul-abil-ir-im 19 İnşallah more after work-1SG -DAT be.useful-SBJP one-thing-PL find-ABIL -AOR -1SG ‘God willing, I will be able to find something useful for my work later.’

(42) İnşallah 2015-te de beraber ol-ur-uz, sayın arkada ş-ım 20 İnşallah 2015-LOC also together be-AOR -1PL dear friend-1SG ‘God willing, we will be together in 2015 too, my dear friend.’

İnşallah, however, is infrequent with the Progressive, and even when it is used it is with the sense of PROGRESSIVE (43), not BOULETIC FUTURE .

(43) İnşallah bu bir ba şlangıç ol-uyor genç arkada ş-lar-ımız için 21 İnşallah this one beginning be-PROG young friend-PL -1PL for ‘God willing, this is becoming a beginning for our young friends.’

This statement, made by Turkish President Abdullah Gül about an increase in the minimum wage, uses the Progressive because the increase had previously been made, so the event of becoming a beginning is already in progress. This is the only possible reading for the Progressive with “İnşallah.”

4.3 Conditional

Another use of the Aorist is for counter-factual statements as a type of CONDITIONAL , which expresses what would happen in the world, if not for another actuality:

(44) Hemen gid-er-im ama karga şa-dan kork-uyor-um 22 Immediately go-AOR -1SG but turmoil-ABL fear-PROG -1SG ‘I would go immediately, but I am afraid of turmoil.’

(45) Yap-ar-ım ama üşen-iyor-um 23 Do-AOR -1SG but be.lazy-PROG -1SG ‘I would do it but I’m lazy.’

19 From a personal website: http://bahadirmemis.com/ 20 From an editorial in Milliyet, a popular newspaper: http://blog.milliyet.com.tr/ataturk-un-agzindan-demokrasi- dersleri/Blog/?BlogNo=333459 21 From a popular news site: http://www.haber365.com/Haber/Abdullah_Gul_Dersim_ Arsivlerinin_ Acilmasinda_Sakinca_Yok/ 22 From a headline about Hosni Mubarek, the former President of Egypt: www.haber365.com/Haber/Hemen_ Giderim_Ama_Kargasadan_Korkuyorum/ 23 A twitter post: http://twitter.com/#!/ecembeceren/statuses/164883131768377345

Caro Semantics of the Turkish Non-Past

(46) İş plan-ı yap-ar-ım fakat hiç tut-maz; iş-ler-im-i hep Work plan-POSS do-AOR -1SG but never hold-NEG.AOR work-PL -1SG -ACC always gecikme-li yap-ar-ım 24 lateness-with do-AOR -1SG ‘I would make a work schedule but it will never hold; I always do my work late.’

This semantic category, which I will call CONDITIONAL , has much in common with the INERTIAL

FUTURE . The INERTIAL FUTURE has a structure which, given the properties of the world, holds

that an event will take place as long as those properties continue to hold. The CONDITIONAL , on the other hand, states that, given a certain set of properties of an individual in both this world and closely related possible worlds, an event would take place, but something in this world is preventing it (while it does occur in those worlds without this or some other preventer). In both cases the use of the Aorist is predicated on the existence of a set of properties which characterize this and other worlds.

4.4 Historical Present

The Aorist is also used for the HISTORICAL PRESENT (or DRAMATIC PRESENT or NARRATIVE

PRESENT ), in which past events are recounted without past-tense morphology for pragmatic reasons.

(47) Kıbrıs, güzel ada, 19. yüzyıl-da uğra ş-ma-ya ba şla-r25 th Cyprus beautiful island 19 century-LOC struggle-FNOM -DAT begin-AOR ‘Cyprus, the beautiful island, begins to struggle in the nineteenth century.’

(48) Limón, 1926-da lise-den mezun ol-ur 26 Limón 1926-LOC high.school-ABL graduate-AOR ‘Limón graduates from high school in 1926.’

(49) Romantizm ate ş-i 20. yüzyıl-da da devam ed-er 27 th Romanticism fire-POSS 20 century-LOC also continue-AOR ‘The flame of romanticism continues in the twentieth century, too.

24 Posted on a forum: http://www.inploid.com/t/is-plani-yaparim-fakat-hic-tutmaz-islerimi-hep-gecikmeli-yaparim- sizce-ne-yapmaliyim/ 25 From a nationalist education charity: http://www.ulkuocaklari.org.tr/bir-gunes-doguyor-sevdamiza-nagehan- penpeci.html 26 From a biography on a dance organization’s website: http://www.bgst.org/dans/arastirma.asp?id=3&bn=1 &righthtml=hlimon-ga0212 27 From a review of a book on a literature site: http://www.sabitfikir.com/elestiri/isyan-ve-melankoli

Caro Semantics of the Turkish Non-Past

Crucially, example (49) was written in 2010, well into the twenty-first century; only if it had been written in 1999 or before would it be possible with the Progressive. It is important to note that all of these examples come from discourse where the Aorist, exclusively, is used to discuss the past. Modern Turkish morphology allows the Aorist, Progressive, and Future affixes to be used in combination with additional Past and Reportative suffixes. Turkish speakers, however, often omit the second suffix until the last verb in series, as in (50-51):

(50) Kız karde ş-im Sıdıka herkes-in muz-u-nu bitir-me-si-ni bekle -r, Sister-1SG Sıdıka everyone-GEN banana-POSS -ACC finish-FNOM -POSS -ACC wait-AOR sonra kendisi-nin-ki-ni ye-me-ye ba şla-r-dı 28 after self-GEN -REL -ACC eat-FNOM -DAT begin-AOR -PST ‘My sister Sıdıka would wait for everyone to finish their banana, then she would begin to eat her own.’

(51) Bir kedi-m ol-sa o-nu ok şa-r, sev -er , o-ndan can al -ır-dı-m29 One cat-1SG be-COND it-ACC pet-AOR love-AOR it-ABL soul take-AOR -PST -1SG ‘If I had had a cat, I would have petted it, loved it, and taken its life.’

The combination of Aorist and Past, however, has a decidedly different meaning than the

HISTORICAL PRESENT . In each of these examples it is one of the uses of uses of the Aorist

combined with the Past: the CHARACTERISTIC in (50) and the counter-factual CONDITIONAL in (51) (because the speaker never did, in fact, have a cat).

4.5 Precative

The Aorist is also used to make polite requests, as in (52-53):

(52) Pencere-yi aç-ar mı-sın? Window-ACC open-AOR Q-2SG ‘Would you open the window?’

(53) Ban-a bir bardak su getir-ir mi-sin? I-DAT one glass water bring-AOR Q-2SG ‘Would you bring me a glass of water?’

28 From an editorial in a popular newspaper: http://haber.gazetevatan.com/aziz-yildirima-istenen- ceza/415334/4/Haber 29 From an article on the childhood of the chief of the Justice Department: http://www.hurriyet.com.tr/gundem /20186238.asp

Caro Semantics of the Turkish Non-Past

These requests cannot be accommodated by the Progressive or Future, which are interpreted as questions about a current action (54a, 55a) or plan (54b, 55b), respectively.

(54) a. Pencere-yi aç-ıyor mu-sun? Window-ACC open-PROG Q-2SG ‘Are you opening the window?’

b. Pencere-yi aç-acak mı-sın? Window-ACC open-FUT Q-2SG ‘Are you going to open the window?’ 30

(55) a. Ban-a bir bardak su getir-iyor mu-sun? I-DAT one glass water bring-PROG Q-2SG ‘Are you bringing me a glass of water?’

b. Ban-a bir bardak su getir-ecek mi-sin? I-DAT one glass water bring-FUT Q-2SG ‘Are you going to bring me a glass of water?’

This use of the Aorist for PRECATIVE may seem at first to be incompatible with the other semantic categories expressed by the Aorist, but if the question particle is seen as scoping over the entire sentence (or at least the verb phrase) then it can easily be interpreted as simply the

interrogative of the INERTIAL FUTURE . That is, if the INERTIAL FUTURE in (56a) makes a statement along the lines of “Given the properties of this world, I (the speaker) expect you to open the window,” then the question in (56b) is simply “Is this world such that, given its properties, I (the speaker) can expect you to open the window?”

(56) a. Pencere-yi aç-ar-sın Window-ACC open-AOR -2SG ‘You’ll open the window.’

b. Pencere-yi aç-ar mı-sın? Window-ACC open-AOR Q-2SG ‘Would you open the window?’

This interpretation may seem farfetched, but polite requests are often quite removed from simple imperatives. 31 Indeed, in the case where the question is less clearly a request, the

30 I have chosen to translate these sentences with the going to futurate because the Future with will can be interpreted as a request in English (e.g. Will you open the window? ) 31 The English expression “Would you be so kind as to…?” comes to mind as an example of this phenomenon.

Caro Semantics of the Turkish Non-Past

interrogative of the Aorist can easily be interpreted as either a request or a question about the properties of the world (i.e. the interrogative of INERTIAL FUTURE ), as in (57).

(57) Hafta-ya New York-a gid-er mi-sin? Week-DAT New York-DAT go-AOR Q-2SG a. POLITE REQUEST : ‘Will you go to New York next week? I need you to buy me some spices…’ b. INERTIAL FUTURE : ‘Will you go to New York next week? All of your friends are going. [Are you the type of person who follows her friends?]’

On the other hand, the Progressive and Future can only express a request for information, rather than action:

(58) a. Hafta-ya New York-a gid-iyor mu-sun? Week-DAT New York-DAT go-PROG Q-2SG ‘Are you going to New York next week?’

b. Hafta-ya New York-a gid-ecek mi-sin? Week-DAT New York-DAT go-FUT Q-2SG ‘Will you go to New York next week?’

4.6 Performative

The final sense of the Aorist is its PERFORMATIVE use, in which a speaker, by uttering a sentence in the Aorist, commits herself (or someone else) to a future action.

(59) In a planning meeting: Afi ş-ler-i ben as-ar-ım Poster-PL -ACC I hang-AOR -1SG ‘I’ll put up the posters.’

(60) Dur, anne, ben yap-ar-ım 32 Stop mother I do-AOR -1sg ‘Stop, mother; I’ll do it.’

(61) Siz yorul-ma-yın, ben market-e gid-er-im 33 You tire-NEG -2PL I market-DAT go-AOR -1SG ‘Don’t tire yourselves; I’ll go to the market.’

32 The title of a blog post with pictures of the author’s child: http://samifarag.blogspot.com/2008/02/dur-anne-ben- yaparim-demeye-basladik.html 33 From a forum on chivalry and sexism: http://www.komikler.com/komikforum/baslik.php?tid=15480&p=2

Caro Semantics of the Turkish Non-Past

This performative use is problematic in that it appears to be bouletic, as the speaker is committing someone to be the director of the action and to see that act through. This issue is taken up again in section 7, below, where the distinctions between the Aorist, Progressive, and

Future with PERFORMATIVE functions are discussed.

5. Uses of the Future

The Turkish Future takes the form of a suffix –AcAk . It is a prototypical future tense which, as Dahl (1985) points out, does makes no requirements of intentionality; rather, its only requirement is that of Futue Time Reference.

5.1 Future Time Reference

Just as in English (see (13), above), the Future in Turkish can have both bouletic and inertial readings. While Yava ş categorizes the Turkish Future as being indicative of some intermediate level of certainty between the Aorist and Progressive, in actuality it can be applied to any statement with future reference, including those cases in which the Aorist or Progressive are applicable, as in (62).

(62) a. Karar ver-di-k, sene-ye Amazon-a gid-ece ğ-iz Decision give-PST -1PL year-DAT Amazon-DAT go-FUT -1PL ‘We’ve decided, next year we’ll go to the Amazon.’

b. Amazon çok enteresan ol-malı; henüz gör-me-di-k, herhalde sene-ye gid-ece ğ-iz Amazon very interesting be-NEC yet see-NEG -PST -1PL probably year-DAT go-FUT -1PL ‘The Amazon must be very interesting; we haven’t seen it yet; we’ll probably go next year.’

Indeed, the Future in Turkish, as in English, can be used with future reference for statements which are entirely uncertain and over which the speaker has no control, as in (63), and in those which are entirely certain, as in (64). 34

(63) Yarın Fenerbahçe Galatasaray-ı yen-ecek Tomorrow Fenerbahçe Galatasaray-ACC beat-FUT ‘Galatasaray will beat Fenerbahçe tomorrow.’

34 If any event in the future is, in fact, entirely certain. But to the extent that they are certain, Turkish speakers can use the Future to talk about them.

Caro Semantics of the Turkish Non-Past

(64) Sene-ye ya ş-ım 21 ol-acak Year-DAT year-1SG 21 be-FUT ‘Next year I will be 21 years old.’ (lit. ‘Next year my age will be 21.’)

Additionally, the tests used above to determine whether statements were bouletic or inertial always succeed with the future; it is compatible with “ İnşallah,” a sign of inertial ordering, as in (65-67), and with phrases like ‘I have decided’, which require bouletic ordering, as in (68-69).

(65) Türkiye, sigara-dan kendi halk-ı-nı koruma sıralama-sı-nda Turkey cigarette-ABL self people-POSS -ACC protection ranking-POSS -LOC in şallah dünya birinci-si ol-acak 35 in şallah world first-POSS be-FUT ‘God willing, Turkey will be first in the world ranking in protecting its own people from smoking.’

(66) Ben muhabbet-ler-imiz-e bugün de ğil de in şallah yarın yaz-aca ğ-ım 36 I chatter-PL -1PL -DAT today not also in şallah tomorrow write-FUT -1SG ‘God willing, if not today then tomorrow I will write about our chatter.’

(67) En geç yarın-a kadar in şallah bir çözum bul-aca ğ-ız 37 Most late tomorrow-DAT until in şallah one solution find-FUT -1PL ‘God willing, we will find a solution by tomorrow at the latest.’

(68) Ama karar ver-di-m küçük de ol-sa yap-aca ğ-ım 38 But decision give-PST -1SG small also be-COND do-FUT -1SG ‘But I’ve decided, I will do it even if it is small.’

(69) Hafta-ya takım-la beraber antrenman-lar-a ba şla-yaca ğ-ım 39 Week-DAT team-with together training-PL -DAT begin-FUT -1SG ‘Next week I will start training together with the team.’

These examples demonstrate that the Future can be used with all statements with Future Time Reference, regardless of whether they have inertial or bouletic ordering.

35 From a news article: http://www.ntvmsnbc.com/id/25336957/ 36 From an editorial on a sports news site: http://www.sporexpres.com/bu-maya-tutar.html 37 From the comments on an article about a new television series: http://www.jyjturkey.com/rooftop-prince-dizisi-1- bolum-turkce-altyazisi-vikiye-eklenmistir.html 38 From an environmentalist forum: http://www.agaclar.net/forum/uyelerin-bahceleri/18797-97.htm 39 From an article about a women’s soccer team in a regional newspaper: http://www.cerkezkoyhaber.com.tr/haber/10638/galibiyet-icin-yola-ciktilar.html

Caro Semantics of the Turkish Non-Past

5.2 Performative

The Future, like the Aorist, also has a PERFORMATIVE use, and the examples from section 0.0, above, are equally acceptable with the Future.

(70) In a planning meeting: Afi ş-ler-i ben as-aca ğ-ım Poster-PL -ACC I hang-FUT -1SG ‘I’ll put up the posters.’

(71) Dur, anne, ben yap-aca ğ-ım Stop mother I do-AOR -1sg ‘Stop, mother; I’ll do it.’

(72) Siz yorul-ma-yın, ben market-e gid-ece ğ-im You tire-NEG -2PL I market-DAT go-AOR -1SG ‘Don’t tire yourselves; I’ll go to the market.’

All performative statements refer, perforce, to a point in time after speech time. Because the Future is used for all statements with Future Time Reference, as demonstrated above, this performative use is much less surprising than the Aorist with the same meaning.

6. Formalizing the Distinction between the Aorist and Progressive

The basic distinction between the Aorist and Progressive is that the Aorist denotes a property of an individual, regardless of time, while the Progressive can only be used to report the properties of a specific period of time which includes the time of utterance (though this time period can be unbounded). This was clear in examples (24) and (25), repeated here as (73) and (74), in addition to (75).

(Yava ş 1982) (73) Baba-m erken yat-ar ama son iki sene-dir geç yat-ıyor Father-1SG early lie.down-AOR but last two year-for late lie.down-PROG ‘My father goes to bed early, but these last two years he has been going to bed late’

(74) a. Benim kasab-ım iyi et sat-ar (Yava ş 1982) My butcher-1SG good meat sell-AOR ‘My butcher sells good meat.’

Caro Semantics of the Turkish Non-Past

b. Benim kasab-ım iyi et sat-iyor My butcher-1SG good meat sell-PROG ‘My butcher is selling good meat.’

(75) Migros normal-de pahalı sat-ar ama bu hafta ucuz-a sat-ıyor Migros normal-LOC expensive sell-AOR but this week cheap-DAT sell-PROG ‘Migros normally sells [this] expensive, but is selling [it] for cheap this week’

In (73), going to bed early is a characteristic of my father, while going to be bed late is merely a property of the time period which includes the utterance time. Similarly, in (74a) selling good meat is a characteristic of my butcher, while in (74b) my butcher selling good meat is merely a property of a time period which includes now. The distinction between the Aorist and Progressive holds even when these grams are used in combination with other morphemes, such as the negative, interrogative, and abilitative infixes, shown below in (76-79).

(76) a. Rakibe Ankara-ya git-mez Rakibe Ankara-DAT go-NEG .AOR ‘Rakibe does not go to Ankara [ever].”

b. Rakibe Ankara-ya git-m-iyor Rakibe Ankara-DAT go-NEG -PROG ‘Rakibe is not going to Ankara [now].”

In (76a) not going to Ankara is a characteristic of Rakibe, whereas in (76b) Rakibe not going to Ankara is simply a situation which holds at a time that includes now.

(77) a. Ali protesto yap-ar mı? Ali protest do-AOR -Q ‘Does/would Ali [ever] protest?’

b. Ali protesto yap-ıyor mu? Ali protest do-PROG -Q ‘Is Ali protesting?’

Similarly with the interrogative affix, the Aorist question (77a) asks “Is it a characteristic of Ali that he would protest, given some specific situation in the world?” while the Progressive

question (77b) asks “At the time NOW , is Ali protesting?”

Caro Semantics of the Turkish Non-Past

(78) a. Baklava-yı yi-yebil-ir-im Baklava-ACC eat-ABIL -AOR -1SG ‘I can eat the baklava ( e.g. because I am not allergic to it)’

b. Baklava-yı yi-yebil-iyor-um Baklava-ACC eat-ABIL -PROG -1SG ‘I can eat the baklava ( e.g. because my diet is over or has not yet started)’

The differences between the Aorist and Progressive still hold with the abilitative suffix in (78); (78a) reports that the speaker has the characteristic of being able to eat baklava in general, while

(78b) only scopes over a certain period of time including NOW . These distinctions hold when more than one morpheme is used in combination with the Aorist and Progressive, as in (79):

(79) a. Esma kayak yap-a-maz mı? Esma ski do-ABIL -NEG .AOR -Q “Can Esma not ski [ever]?”

b. Esma kayak yap-a-m-ıyor mu? Esma ski do-ABIL -NEG -PROG -Q “Can Esma not ski [at the moment]?”

The first question (79a) could be asked, e.g., if Esma had some injury which the speaker believed might prevent her from ever being able to ski, while in (79b) the speaker only wishes to know if something (e.g. a particularly busy schedule or an overbearing parent) is preventing Esma from being able to ski at the current moment.

7. Formalizing the Performative

As mentioned in the first discussion of the Aorist with performative meaning, above

(section 4.6), the paradigm of Progressive with BOULETIC FUTURE and Aorist with INERTIAL

FUTURE encounters a snag in performative statements. Examples, such as (80), show that only the Aorist can be performative in this sense:

(80) In a planning meeting: a. Afi ş-ler-i ben as-ar-ım Poster-PL -ACC I hang-AOR -1SG ‘I’ll put up the posters.’

Caro Semantics of the Turkish Non-Past

b. Afi ş-ler-i ben as-ıyor-um Poster-PL -ACC I hang-PROG -1SG ‘I am (the one who is) putting up the posters.’ 40

Only in sentence (80a) is the speaker announcing that she has volunteered to undertake the task of hanging the posters; that is, only in (80a) is the speaker changing the context in a performative way as outlined by Condoravdi and Lauer (2011). The only interpretation of (80b) is that at some point in the past the speaker (or others) had already decided that she would hang the posters; she is merely reporting information about the current state of the world (without regard to whether any of the listeners already knew this particular piece of information). The Future, however, can be used with both a positive and negative performative sense. (81a,b) are nearly synonymous in their performative meanings; the only distinction is that the Future sentence is perceived as slightly more formal:

(81) a. Ben hoca-yla konu ş-up on-dan ek sure iste-r-im I teacher-with speak-CONJ she-ABL additional period want-AOR -1SG ‘I’ll talk to the teacher and ask her for an extension”

b. Ben hoca-yla konu ş-up on-dan ek sure iste-yece ğ-im I teacher-with speak-CONJ she-ABL additional period want-FUT -1SG ‘I’ll talk to the teacher and ask her for an extension”

How, then, can the Aorist be used performatively when the Progressive cannot? This

question is particularly vexing because the PERFORMATIVE seems to require the speaker to commit someone to being a director, a situation required for the bouletic PROGRESSIVE with

Future Time Reference but incompatible with the AORIST with Future Time Reference.

7.1 The Absence of a Progressive Performative

The Progressive cannot be performative (i.e. cannot change the context in which the sentence is uttered) because the Progressive is limited to reporting a situation which holds at a

time that includes the “ NOW ” of speech time. That is, the event time must have already begun when the speech act occurs. However, in performative statements the event must begin at, or

later than, the time of speech. This crucial fact of the PERFORMATIVE , that the event can only

40 Here the pronoun ben ‘I’ immediately precedes the verb because of focus restrictions in Turkish. In a conversation about hanging posters the critical element of these statements is the subject, i.e. the poster-hanger, so word order changes from the expected SOV to OSV. In these examples the poster-hanging will happen in any case, while the poster-hanger herself is critical, so the English translations could use a semi-cleft construction.

Caro Semantics of the Turkish Non-Past

begin at (or after) the time of the speech act, explains why the Progressive cannot be used to express PERFORMATIVE meanings. The requirement of the Progressive to have the event time begin before speech time, however, does not seem compatible with the many examples of Progressive with Future Time Reference. Indeed, many of the sentences in 3.3 include actions that will occur long after the speech act. In all of these cases, however, there is a plan already in place at the time of speech (Deo 2012). The “planning mechanism”, then, requires someone who has decided to ensure a certain outcome, i.e., a bouletic director. This fact of the Progressive explains its use in the

BOULETIC FUTURE as well as its unavailability for performative statements.

7.2 The Performative Future

The Future can be performative because any performative statement is, perforce, a statement which references future time. Since the Future encompasses all statements with Future Time Reference, it must also be able to be used to encode performative statements.

7.3 The Performative Aorist

Section 0.0, above, demonstrated that the Progressive cannot be used performatively because it requires that the event in question already have started when the statement is made, which cannot be true of performative statements. The performative Aorist is reminiscent of the

PRECATIVE use, which also commits (or, at least, attempts to commit) the addressee, rather than the speaker, to an action. It is also true that the Aorist was, historically, a “Present” before the development of the Progressive in the seventeenth and eighteenth centuries (Johanson 1989). Before this time, the Aorist would have encompassed all of the current uses of the Progressive, including the BOULETIC FUTURE and related PERFORMATIVE . The simplest explanation, then, is

that the Aorist with PERFORMATIVE meaning is a remnant of this history. Still, the semantic map

implied by this explanation does not require a “doughnut,” as the PERFORMATIVE is connected to both the PRECATIVE and the BOULETIC FUTURE .

8. Conclusion

The analysis above leads to the following semantic map, which lays out the relationships between semantic notions and the grammatical morphemes which encompass them.

Caro Semantics of the Turkish Non-Past

Figure 2 . A possible semantic map of the Turkish non-past

This essay has laid out the semantic distinctions between three Turkish grams: the

Progressive, Aorist, and Future. In demonstrated the distinction between the HABITUAL and

CHARACTERISTIC /GENERIC as one of characteristics of time periods and individuals, respectively. It also formalized the distinction in Future Time Reference through Copley’s (2009) bouletic and inertial futures. Finally, it has explained the unexpected use of the Aorist, but not the Progressive, with performative statements, leaving a semantic map in the tradition of Haspelmath (2003).

Caro Semantics of the Turkish Non-Past

References

Bybee, Joan L., and Östen Dahl. 1989. “The Creation of Tense and Aspect Systems in the Languages of the World.” Studies in Language 13.1, 51-103.

Bybee, Joan L., Revere Perkins, and William Pagliuca. 1994. The evolution of grammar: Tense, aspect and modality in the languages of the world . Chicago: The University of Chicago Press.

Comrie, Bernard. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related Problems . Cambridge: Cambridge University Press.

Copley, Bridget. 2009. The Semantics of the Future . New York: Routledge.

Dahl, Östen. 1985. Tense and Aspect Systems . Oxford: Basil Blackwell.

Deo, Ashwini. 2012. “Settling the future: aspectual marking and performative effects in Marathi.” Unpublished abstract.

Dowty, David R. 1977. “Towards a semantic analysis of verb aspect and the English ‘imperfective’ progressive.” Linguistics and Philosophy 1, 45-77.

Dowty, David R. 1979. Word meaning and Montague Grammar . Dordrecht: Reidel.

de Haan, Ferdinand. 2009. “Building a semantic map: top-down versus bottom-up approaches.” Linguistic Discovery .

Fenn, Peter. 1986. A Semantic and Pragmatic Examination of the English Perfect . Tubingen: G. Narr.

Haspelmath, Martin. 1998. “The Semantic Development of Old Presents: New Futures and Subjunctives without Grammaticalization.” Diachronica XV.1, 29-62.

Haspelmath, Martin. 2003. “The Geometry of Grammatical Meaning: Semantic Maps and Cross- Linguistic Comparison.” In Michael Tomarello (ed.), The New Psychology of Language, vol. 2 . Mahwah, New Jersey: Erlbaum, 211-242.

Huddleston, Rodney and Geoffrey Pullum. 2002. The Cambridge Grammar of the English Language . Cambridge: Cambridge University Press.

Jaszczolt, Katarzyna M. 2003. “The modality of the future: A Default-Semantics account’. In P. Dekker and R. van Rooy (eds.), Proceedings of the 14th Amsterdam Colloquium, ILLC ,University of Amsterdam, 43-48.

Johanson, Lars. 1989. “Aorist and Present in West Oghuz Turkic.” Journal of Turkish Studies 13, 99-105.

Caro Semantics of the Turkish Non-Past

Johanson, Lars. 2009. “Türk Dillerinde Odaksıl Şimdiki Zaman İş aretleyicisi ve Kıbrıs Türkçesinde Eksikli ği.” Bahar 49, 93-100.

Kemiro ğlu, Caner. 2010. “On the Epistemic Modality in Turkey Turkish: Uncertainty.” Turkish Studies 5.4, 434-478.

Kornfilt, Jaklin. 1997. Turkish . New York: Routledge.

Leech, Geoffrey N. 1971. Meaning and the English Verb . London: Longman.

Lewis, Geoffrey. 2000. Turkish Grammar . Oxford: Oxford University Press.

Malchukov, Andrej L. 2010. “Analyzing Semantic Maps: A Multifactorial Approach.” Linguistic Discovery 8.1, 176-198.

Menges, K. 1968. The Turkic Languages and Peoples . Weisbaden: Ural Altaische Bibliothek.

Searle, John R. 1975. “A taxonomy of illocutionary acts.” In K. Gunderson, ed., Language, Mind, and Knowledge . Minneapolis: University of Minnesota Press.

Taylan, Eser Erguvanlı, ed.. 2001. The Verb in Turkish . Philadelphia: J. Benjamins. von Fintel, Kai. 2006. “Modality and Language”. In Donald M. Borchert (ed.), Encyclopedia of Philosophy – Second Edition . Detroit: MacMillan Reference.

Yava ş, Feryal. 1982. “The Turkish Aorist.” Glossa 16, 40-53.

Vendler, Zeno. 1957. “Verbs and Times.” The Philosophical Review 66, 143-160.